The recent developments in the research area of electric vehicle (EV) dynamics and energy management are significantly advancing the field through innovative approaches that integrate machine learning with physical models. A notable trend is the use of Physics-Informed Neural Networks (PINNs) to predict EV dynamics and energy consumption, which allows for more accurate path planning and battery usage estimation without the need for additional sensors. This approach leverages real-world parameters and ensures consistency with ground truth data, demonstrating high accuracy and generalization capabilities. Another significant advancement is the optimization of energy dispatch for grid-connected EVs, where refined battery models, such as the electrochemical model, are being integrated into the dispatch process to enhance efficiency and prevent battery degradation. This involves complex optimization techniques that manage the state distribution of charging EVs and address computational challenges associated with mixed integer optimization. Additionally, there is a growing focus on learning-based system identification for autonomous racing vehicles, where neural networks are employed to improve tire modeling accuracy in dynamic racing conditions. This method allows for rapid and accurate identification of tire parameters directly on the race track, overcoming the limitations of traditional methods that require extensive operational ranges or specific experimental setups. These innovations collectively push the boundaries of EV technology and autonomous racing, offering more precise control and optimization strategies.